Agents of change: What is an AI agent and how can it help your underwriting business evolve?

After years of mounting debate and discussion, 2025 must be the year AI talk turns to action in the insurance industry. But how do commercial and specialty insurers start tackling AI without becoming overwhelmed and tangled up in grand plans for change?

We hear it all the time. Insurers know they need to explore AI and digitise their business, but don’t know where to start. A full overhaul would be excruciatingly expensive, resource-heavy and risk alienating hard-working employees who feel their roles are being stripped out. No two businesses are at the same point in their digital journey, so enhancing processes via AI may mean very different things to different people.

We believe thoughtful, intentional and incremental change is far more effective in the long term than sweeping change. These small but meaningful changes will be driven by AI agents.

We’re going to hear a lot of talk about AI agents this year. Forbes recently published an article explaining the concept, and Google listed AI agents as one of its top trends in AI in 2025.

AI agents are the vehicle through which talk of AI turns to action. We’re now getting to the heart of how AI can be practically deployed to move the digitisation dial.

Enter the agents

Think of an AI agent as an AI-powered app designed to automate and streamline a specific part of the underwriting process, for example, assessing risk appetite, or producing a renewal comparison. Each agent is programmed to automate this one process; it’s their only job, and they do it well.

Strong, profitable underwriting is a sum of its parts. Each stage of the process involves lots of individual moments, points at which underwriters have to make choices based on the information they’re presented with. Some of these moments require the finely honed skill of years of expertise, but many don’t.

Most underwriters we speak to find that the process of binding a risk is full of moments of tedium, researching a company or location, trying to gauge public sentiments about a business, and scanning submissions to assess whether they fit their risk appetite. These processes are at the heart of underwriting, but they’re not the soul.

Less talk, more action

The beauty (and challenge) of today’s open AI platforms is that anyone can use them to build apps to support their business requirements. There are practically no limits to what AI could be able to automate now and in the future.

This is great news for insurers. Rather than tackling AI as a nebulous whole, they can focus on the individual areas where automation would be most beneficial, and deploy AI agents to tackle those discrete parts.

Whether building their own or buying out-of-the-box AI solutions from an expert provider, AI agents are ready to revolutionise your underwriting business, one step at a time.

We’d love to talk about how Send can support the build and integration of AI agents to increase your underwriting efficiency. Get in touch here.

Matt McGrillis, co-founder and Chief Product Officer, Send